Osteoarthritis (OA) of the knee and hip joints is a debilitating condition that has a high prevalence worldwide. While the disease is not fatal, it has a significant impact on joint function resulting in pain and reduced range of motion (1). In the early stages of OA, conservative treatment modalities like activity modification, physical therapy, exercise are recommended to delay disease progression, however, they are often not effective in providing rapid relief in patients with symptomatic OA (2).

In patients with symptomatic and end-stage OA, intra-articular injections are frequently administered for both diagnostic and therapeutic purposes (3–5). Such injections are helpful to delineate the cause of pain arising from the intra-articular source or due to extra-articular causes like the spine or surrounding musculature (6). Furthermore, several high-quality randomized controlled trials (RCTs) have demonstrated the efficacy of intra-articular injections of corticosteroids (CS) or hyaluronic acid (HA) in managing OA of the knee and hip joints (3, 7). These injections can significantly alleviate pain and improve symptoms in short term thereby delaying the need for total joint arthroplasty (TJA). Indeed, guidelines of the American College of Rheumatology conditionally advocate the use of CS injections for managing OA while the Osteoarthritis Research Society International suggests the use of intra-articular CS for symptomatic OA not responsive to anti-inflammatory drugs and those of signs of local inflammation (8, 9). Despite the results from several RCTs, the use of intra-articular injections has been controversial as well. The use of intra-articular drugs, especially CS, has been associated with rapid progression of osteoarthritis, subchondral insufficiency fracture, osteonecrosis, and rapid joint destruction with bone loss (10). Several other local complications like skin and fat atrophy, septic arthritis, and prosthetic joint infections (PJI) have also been related to the use of intra-articular injections (5, 11).

In the past decade, several systematic reviews and meta-analyses have dwelled on the relationship between the use of prior intra-articular injections and the risk of PJI after TJA (11–15). However, these reviews had several limitations. Foremost, the number of studies included was not high with a maximum of eight comparative studies pooled in the meta-analysis. Furthermore, none of the past reviews took into account the timing of injection and the risk of PJI. Given these shortcomings, there is a need for an updated and comprehensive review on this subject. Therefore, the current review was designed to assess the impact of prior intra-articular injections on the risk of PJI in patients undergoing TJA with a focus on the timing of injection before surgery.

PJI is a debilitating complication seen in 1–2% of patients undergoing TJA and accounts for around 15% of total revision procedures (37). Notwithstanding the intense research in the prevention and management of PJI, there seems to be no decline in the incidence of this devastating complication (38, 39). Prediction models indicate that the incidence of PJI for knee and hip joints is on the rise and is expected to grow by 170 and 176% respectively (39). Amongst the several risk factors identified for PJI, the role of prior intra-articular injections has been controversial. Despite several studies and systematic reviews, it is still unclear how do prior CS and HA injections impact the incidence of PJI (11–15).

The previous meta-analysis on this subject with a small number of studies has produced contrasting results. Charalambous et al. (15) in a pooled analysis of eight studies reported no significant increase in the risk of PJI with prior intra-articular CS injections (RR 1.87; 95% CI 0.80–4.35). On the other hand, Xing et al. (14) demonstrated a significantly higher risk of PJI in patients with prior history of intra-articular CS injections [Odds ratio (OR) 2.13, 95% CI 1.02–4.45]. Li et al. (13) in a review of systematic reviews pointed out the scarcity of evidence and the need for further studies to clarify the role of prior injections on the risk of PJI. In this context, our review is a significant update of prior studies by providing cumulative evidence from 19 studies with 521,267 participants. Our results demonstrated that patients with prior history of intra-articular injections have a statistically significant 24% increased risk of PJI. Overall, the risk of PJI in the injection group was 1.6% while in the control group it was 1.47%. While these figures are within the range reported in the literature (37), the significant increased incidence of PJI with prior intra-articular injections cannot be dismissed. PJI is a catastrophic complication with huge financial burden, morbidity and mortality (38, 39); and the small increased risk needs to be interpreted in the magnitude of the potentially preventable complication.

Recently, it has been suggested that the timing of injection before surgery is an important variable while assessing the risk of PJI (30). Indeed, since the pathophysiological mechanism for the heightened risk of infection with prior injections is based on the prolonged immunosuppressive effects of CS and the joint contamination caused by the injection procedure (15), it would be plausible to believe that reduced time interval between injection and surgery would increase the risk of PJI. To explore such a relationship, we performed multiple subgroup analyses based on the timing of injections but with contrasting results. In the largest subgroup of studies with <12 months' time interval between injections and surgery, we noted a statistically significant 10% increased risk of PJI in the injection group. However, on a meta-analysis of studies with a time interval of <6 months, 3–6 months, 0–3 months, and <1month between injection and surgery, we noted no significant increase in the risk of PJI. Nevertheless, on careful examination of the effect size and 95% CIs, it can be noted that there was a tendency of increased risk of PJI in the injection cohort in all these subgroups but the difference could not reach statistical significance. One reason for this could be the limited number of studies in each analysis which reduced the statistical power. Since the risk of PJI is only marginally increased with prior intra-articular injections, the number of participants in the injection cohort of these subgroups may not have been sufficient to produce statistically significant results.

The majority of studies in our review did not present separate data for CS and HA which limited our ability to differentiate the risk of PJI with these two drugs. In a small subgroup analysis of five studies with <12 months' time interval between injections and surgery, we noted no significant impact of only CS injections on the risk of PJI. To date, very few studies have compared the risk of PJI with prior HA and CS. Kurtz et al. (31) in a large retrospective review have noted no difference in the risk of PJI with prior CS or HA injections. Similarly, Amin et al. (22) also found no difference between the two drugs while assessing the risk of PJI with prior intra-articular injections. Given the scarcity of data, further studies comparing the two drugs for the risk of subsequent PJI are needed to derive stronger conclusions.

Another factor not considered in our results is the number of injections before surgery. However, a few studies have examined this issue. In one of the excluded studies, Richardson et al. (40) compared the risk of PJI after TKA between those receiving multiple injections vs. those receiving single intra-articular injections before surgery and found no difference in the risk of PJI. On the other hand, Forlenza et al. (30) have demonstrated a dose-dependent relationship between the number of injections and risk of PJI after THA with every unit increase in the number of injections increasing the risk of infections. These contrasting results between THA and TKA are difficult to explain especially when our meta-analysis has failed to demonstrate significant differences in complication rates between THA and TKA. In another study, Kokubun et al. (41) have compared infection rates in TKA patients receiving ≤ 3 vs. >4 injections before surgery and found no difference between the two groups. Given such heterogeneous results, the exact relationship between the number of injections and the risk of PJI too deserves further research.

The increased risk of PJI with prior intra-articular injections seen in our meta-analysis should be interpreted with the fact that several confounders can influence this outcome. Research has shown that comorbidities like diabetes, obesity, malnutrition, smoking, tobacco use, narcotic use, alcohol dependence, and methicillin-resistant Staphylococcus aureus colonization of the nares can increase the risk of PJI (42). Tang et al. (19) in their study assessing predictive factors for use of intra-articular injections before TJA have demonstrated that less healthy patients who eventually require TJA more often undergo intra-articular injections to delay surgery as compared to healthier patients. Therefore, it is plausible that baseline differences between the injection and control groups could contribute to the difference In the risk of infection. To nullify this effect, baseline matching and multivariate analysis are essential. While matched cohorts were used in a few studies, multivariable-adjusted data were not reported by majority studies, prohibiting a pooled analysis. In the few studies reporting multivariable-adjusted results, some demonstrated no increased risk of PJI (19, 31) while others demonstrated an increased risk (21, 34).

The limitations of our review need to be mentioned. Foremost, our analysis is based on data mostly from retrospective cohort studies which have an inherent bias. There would have been obvious selection bias between the injection and control groups which could have skewed the results. Secondly, the majority of our studies were from administrative databases, and data was collected using current procedural terminology (CPT) codes. It is known that such databases are prone to coding errors. Furthermore, since the majority of studies were from databases in the USA, despite taking care to avoid overlapping studies, we may have inadvertently repeated the same patients. Thirdly, all of the studies were from North America and Europe and this limits the generalizability of our results to the global population. Fourthly, due to a lack of data, we were unable to assess the impact of the number of injections and type of drug (CS or HA) on the study outcomes. Due to the same reason, we were unable to pool multivariable-adjusted ratios for the risk of PJI. Lastly, the definition of PJI was either not reported or were varied in the included studies. This has important implications as lack of standardized definition may have overestimated or underestimated the incidence of PJI in the included studies. Also, there was heterogeneity in the duration of follow-up ranging from just 3 months to up to 7 years.

To conclude, the results of our updated systematic review and meta-analysis indicate that patients with prior intra-articular injections have a small but statistically significant increased risk of PJI after TJA. Considering that PJI is a catastrophic complication with huge financial burden, morbidity and mortality; the clinical significance of this small risk cannot be dismissed. The question of the timing of injections and the risk of PJI still remains and can have a significant impact on the decision making. Further high-quality data is needed focussing especially on the timing, type and number of injections and the risk of PJI after TJA. Future studies should also use standardized definitions of PJI to allow comparisons between different cohorts.

The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author/s.

FN and WL conceived and designed the study. FN was involved in literature search and data collection and wrote the paper. WL analyzed the data and reviewed and edited the manuscript. All authors read and approved the final manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.